The long-term objectives of this research proposal are to understand the social and contextual factors that influence perinatal health and preterm birth, beyond individual risk factors and behaviors. Previous studies have investigated racial discrimination on the individual level yet few studies explore institutional measures of racism as potential social, contextual factors in association with the racial and ethnic disparities in preterm birth. This research project will apply multilevel methods to investigate both individual-level measures of discrimination and institutional racism measured in the form of "redlining" (housing discrimination based on race) and residential segregation in association with maternal stress and preterm birth among pregnant women. The specific aims are to: (1) Explore the development of measures of institutional racism in the form of "redlining" and residential segregation by using The Home Mortgage Disclosure Act (HMDA) database and the US census among an urban population of pregnant women to test its association with maternal stress and preterm birth;(2) Understand the relationships between self-reported measures of individual personally-mediated discrimination and institutional measures of racism;and (3) Measure the associations between self-report of personally-mediated discrimination, institutional racism, stress and preterm birth among a group of women in an urban area. The primary hypothesis for the proposed study is that individual and institutional forms of racism are independently and jointly associated with stress and birth preterm birth. To test this hypothesis, a "redlining" measure will be explored from the HMDA database and an index for residential segregation will be explored and created from the US Census. Secondary multilevel logistic regression analyses will be conducted, linking individual-level data about perceived discrimination, maternal stress and other sociodemographic factors from the Stress, Pregnancy and Evaluation Community Project (SPEAC) with community-level data about "redlining" and residential segregation. SPEAC is a cross-sectional study of pregnant women linked to birth records based in Philadelphia, PA originally designed to understand the relationships between stress and bacterial vaginosis. Exploring institutional and structural community factors that influence perinatal health are of great public health importance. The findings from the proposed study will potentially provide recommendations for public health policy and interventions aiming to eliminate racial/ethnic disparities in preterm birth, which will also influence disparities in infant mortality in the US.